Free forever. Predictable as you grow.

No. Hobby is free forever — no card required, no expiry, no waiting. You only enter payment details when you’re ready to upgrade for more credits, scheduled workflows, or external MCP API access.

One credit represents one meaningful agent action: drafting a document, qualifying a lead, running a report, sending a message. Credits are consumed per action, not per token — so the cost is predictable.

A one-time top-up that sits on top of your subscription credits and never expires. Buy one anytime, on any plan including Hobby. Bigger packs come with bigger discounts — Bulk is 50% off Quick.

Set a threshold (e.g. “when I drop below 100 credits, buy a Boost pack”) and your saved card gets charged automatically when you cross it. We add cooldown and daily caps so it can never run away on you, and every charge fires a notification. Dispatch never pauses, you never get a surprise bill.

Agents pause cleanly — no surprise charges. You can buy a credit pack ($10–$500) for an instant top-up, or enable auto-recharge to never pause again. Your existing data is preserved either way.

You save 20% — pay for 10 months and get 12. Same monthly credit allotment and features as the monthly plan, just billed once a year instead of every month.

Every account gets a unique 8-character code. When a friend signs up with your code (via link or by typing it on the signup page) and later upgrades to a paid plan, you both get 600 bonus credits. Capped at 5 successful referrals per account.

Yes, always. Cancel from your dashboard with no questions asked. You keep full access until the end of your billing period.

Yes. Switch to any lower paid tier or cancel back to Hobby anytime from billing. You keep your subscription credits through the end of the current period; pack credits never expire.

Yes — for teams that need volume pricing, SSO, dedicated onboarding, or custom contract terms, reach out and we’ll scope something together.

No. Your data is never used to train any model, ours or otherwise. Harnyss is bring-your-own-key — your API calls go directly to your chosen LLM provider.

Any OpenAI, Anthropic, or compatible API key works today. Support for Gemini, Mistral, and local models via Ollama is in active development.